CBAM benchmarks: how free allocation reduces the certificates on your goods

A CBAM benchmark is a value in tonnes of CO₂e per tonne of good, set by the Commission for each CN code in the table in point 5 of the Annex to Implementing Regulation (EU) 2025/2620, the act on the calculation of the free allocation adjustment. It represents the free allocation an EU installation making the same good would receive under the EU ETS.

The benchmark is not an emissions figure for your plant and it is not a default value. Default values (Implementing Regulation (EU) 2025/2621) are set per country of origin and stand in for missing emissions data. Benchmarks are the same for every country: the Commission's guidance states that, unlike default values, benchmark values apply equally for all countries. What changes the benchmark for a given CN code is the production route, and for a few goods the production year.

The Commission published the benchmark table as an information-only Excel file for the definitive period in February 2026. The legal text is the Annex to IR (EU) 2025/2620.

EU installations in the CBAM sectors receive part of their EU ETS allowances for free. Each installation is split into sub-installations, and each sub-installation receives its EU ETS product benchmark multiplied by its historical production. CBAM was designed so that an importer pays the same carbon cost as an EU producer: emissions minus free allocation.

EU ETS benchmarks do not map cleanly onto CN codes, and they do not account for precursors bought from other plants. So the Commission derived dedicated CBAM benchmarks from the EU ETS benchmarks, adjusted to the CBAM system boundaries, and defined them for every simple and complex CBAM good, for single production processes and for whole production chains.

For electricity (CN 2716 00 00) the free allocation adjustment is zero: electricity producers do not receive free allocation in the EU ETS.

The benchmark is never deducted in full. It is multiplied by the CBAM factor referred to in Article 10a(1a) of the EU ETS Directive, which falls each year as free allocation is phased out for CBAM sectors, and by the cross-sectoral correction factor (CSCF).

Commission Guidance No. 4 lists the CSCF as 1.0 for every year in its table and marks those values as preliminary; values for 2026 to 2030 are set in Commission Implementing Decision (EU) 2026/1862. The practical reading: in 2026 a good is charged on its embedded emissions minus almost its whole benchmark. By 2030 roughly half the benchmark is gone, and from 2034 there is nothing left to deduct.

Article 31 of Regulation (EU) 2023/956 and the Annex to IR (EU) 2025/2620 set the arithmetic. For each imported good g in year y:

FAA(g) = SEFA(g,y) × M(g)

where FAA is the free allocation adjustment, SEFA the specific embedded free allocation in tCO₂e per tonne of good, and M the mass imported in the year the declaration covers.

When actual data is used, the operator calculates SEFA for each production process:

SFA(process) = CBAM factor(y) × CSCF(y) × BM*(g)

For a simple good, SEFA equals that process value. For a complex good, the SEFA of each precursor is added, weighted by the specific mass of precursor used per tonne of good:

SEFA(g,y) = SFA(process) + Σ m(i,y) × SEFA(i)

Precursors from the EU, and from the countries exempted in point 1 of Annex III to the CBAM Regulation, count as zero free allocation, just as their embedded emissions do.

The adjustment then enters the number of certificates the declarant surrenders: embedded emissions minus the free allocation adjustment, minus any eligible carbon price paid in the country of production, with a floor of zero. The full formula is on the CBAM certificate price page.

The benchmark table in IR (EU) 2025/2620 has two columns, and choosing the wrong one is the most common error:

• Column A (BM*) is the benchmark for the production process alone. It applies when SEFA is built from actual data, process by process, with precursors added on top.
• Column B (BM) covers the whole production chain, precursors already included. It applies when the declaration uses default values for the good.

For a simple good the two columns are equal. For a finished steel bar they are far apart: rolling alone carries a small benchmark, while the chain from iron ore or scrap to bar carries the large one.

Some CN codes carry several values marked with a letter. For steel the letter is the production route: (C) carbon steel from blast furnace and basic oxygen furnace, (D) carbon steel from direct reduced iron and electric arc furnace, (E) carbon steel from scrap and electric arc furnace, and (F), (G), (H) for the same three routes in low alloy steel, with (J) for high alloy steel. Aluminium is split into (K) primary and (L) secondary aluminium, and cement clinker into grey and white clinker. A few values depend on the production year instead: (1) for 2026 to 2027 and (2) for 2028 to 2030.

If the declarant has no verified actual data and uses a default value for the good's embedded emissions, the free allocation adjustment must also be calculated on defaults: SEFA = CBAM factor × CSCF × the Column B benchmark.

The declarant usually does not know the production route of a plant it has never seen. The rules solve that by reading the route from the default value tables in Annex I to IR (EU) 2025/2621: the last column of each country table names the production route used to determine the benchmark. If the country of origin is unknown, the route comes from the table for other countries and territories, or from Annex IV for a precursor.

So a default declaration fixes two things at once: a marked-up default emissions value and a benchmark based on the route the Commission assumed for that country. An installation on a lower-emitting route than the one assumed gets no credit for it unless actual data is supplied.

Only the authorised declarant calculates the final free allocation adjustment. But when actual values are used, the inputs come from the installation, and IR (EU) 2025/2547 and the Free Allocation Adjustment Act set what they must contain. The Commission's guidance states that the verified operator's emissions report contains all necessary information on SEFA. In practice that means:

• the CN code and the production process each good belongs to, with its production route where the benchmark depends on it;
• the production year of the goods, which decides the year-dependent benchmarks and the CBAM factor;
• for complex goods, the mass of each precursor consumed per tonne of good, its country of origin and the SEFA from the supplier's verification report, or the default where none exists;
• a verification report from an accredited verifier covering the reporting period, without which the declarant cannot use the actual values.

A plant that monitors these items well gives its importer the lower of two numbers twice: actual embedded emissions instead of a marked-up default, and the benchmark of its real route instead of an assumed one.